Multivariate patterns from Silvestrini et al.
Citation:
Images
This repository includes 3 images. Each is a map trained using a Support Vector Machine (SVM) to be sensitive to Stroop task demand (incompatible - compatible) or shock pain (pain - rest) within a pre-defined mask.
We used linear support vector machines (SVMs) (Vapnik, 2013) to train multivariate pattern classifiers within aMCC for the task and the shocks. The task-related aMCC mask corresponded to an a priori meta-analytic map including aMCC activity induced by pain, cognitive control, and negative affect (Shackman et al., 2011) (data retrieved from http://neurovault.org). The shock-related aMCC mask consisted of the conjunction of the previous meta-analytic map with the NPS to obtain an aMCC subregion of the NPS. The SVMs were implemented using custom Matlab code based on the Spider toolbox (http://people.kyb.tuebingen.mpg.de/spider). The pattern classifiers were trained on first-level contrast images for the two experimental conditions to discriminate the high and the low demand condition during the task and during the shocks.
dACC_task_pattern_wani_121416.nii
Voxels masked to include dACC only, then trained using SVM to identify pattern weights related to Stroop task demand.
dACC_pain_pattern_wani_121416.nii
Voxels masked to include dACC only, then trained using SVM to identify pattern weights related to shock.
stroop_pattern_wani_121416.nii
Voxels masked to reflect Stroop-related activation in Neurosynth (stroop_pFgA_z_FDR_0.01.nii) then trained using SVM to identify pattern weights related to Stroop task demand.
Loading/using images
You can load the images directly by name using any of the major imaging software packages, or the fmri_data object class in CANLab Core Tools. See below for an example.
There's also an option to load the Stroop image this way:
[obj, name] = load_image_set('stroop');
Defaults settings have been modified by file(s):
/Users/torwager/Dropbox (Dartmouth College)/Matlab_code_external/spm12/spm_my_defaults.m
Modified fields: stats
sampleto = '/Users/torwager/Documents/GitHub/Neuroimaging_Pattern_Masks/Multivariate_signature_patterns/2020_Silvestrini_Rainville_Pain_CogControl_interaction_aMCC/stroop_pattern_wani_121416.nii'
Defaults settings have been modified by file(s):
/Users/torwager/Dropbox (Dartmouth College)/Matlab_code_external/spm12/spm_my_defaults.m
Modified fields: stats
Loaded images:
/Users/torwager/Documents/GitHub/Neuroimaging_Pattern_Masks/Multivariate_signature_patterns/2020_Silvestrini_Rainville_Pain_CogControl_interaction_aMCC/stroop_pattern_wani_121416.nii
The image values contain SVM weights, so the most straighforward use is to calculate the dot product between the images and one or more test images. e.g., if you have the Kragel 270 dataset downloaded, this:
test_pattern_on_kragel_2018_n270_data(obj);
Process: test_pattern_on_kragel_2018_n270_data
----------------------------------------------
Loading /Users/torwager/Dropbox (Dartmouth College)/COURSES/Courses_Dartmouth/2021_3_Spring_fMRI_Class/PSYC60_Shared_resources_for_students/datasets/kragel_2018_nat_neurosci_270_subjects_test_images.mat
Warning: Some images have zero values in some of the 22927 voxels in weight mask. These will be excluded from similarity analysis image-wise.
Number of zero or NaN values within weight mask, by input image:
57 57 57 57 57 57 57 57 57 57 57 57 57 57 57 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 35 0 6 0 3 3 0 0 4 58 0 0 0 0 0 42 101 122 57 53 68 85 595 88 75 3 153 42 34 84 0 0 0 0 11040 0 0 0 0 0 0 0 0 0 0 69 26 36 27 74 36 39 59 56 22 9 9 23 42 52 0 0 35 0 0 0 0 0 0 0 0 0 0 0 0 304 304 304 304 304 304 304 304 304 305 304 304 305 304 304 281 549 142 149 311 259 174 373 388 312 312 247 372 559 223 427 442 159 226 319 521 204 381 309 278 160 295 253 213 435 0 0 0 0 0 0 0 0 0 0 0 0 0 381 0 0 0 0 0 0 0 0 0 0 14196 0 12533 0 0 0 591 666 777 248 691 497 459 670 623 509 192 776 474 480 270 61 61 61 66 202 61 373 93 70 61 68 80 61 61 61 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 20 0 0 0
Bootstrapping CIs
----------------------------------------------
Pattern classification results
----------------------------------------------
Applied pattern to test data from 18 studies, n = 270 participants
From Kragel et al. 2018, Nature Neuroscience
Classification of pain vs. other
ROC_PLOT Output: Single-interval, Optimal balanced error rate
Threshold: 0.04 Sens: 47% CI(36%-57%) Spec: 47% CI(39%-54%) PPV: 30% CI(23%-38%) Nonparametric AUC: 0.43 Parametric d_a: -0.26 Accuracy: 47% +- 3.0% (SE), P = 0.300852
Threshold for cosine_sim = 0.0392, Cohen's d(pain vs no) = -0.26, Bootstrapped CI is [-0.50 -0.01]
Classification of cognitive control vs. other
ROC_PLOT Output: Single-interval, Optimal balanced error rate
Threshold: 0.04 Sens: 49% CI(39%-59%) Spec: 49% CI(42%-56%) PPV: 32% CI(25%-40%) Nonparametric AUC: 0.50 Parametric d_a: 0.00 Accuracy: 49% +- 3.0% (SE), P = 0.760972
Threshold for cosine_sim = 0.0399, Cohen's d(cog vs no) = 0.00, Bootstrapped CI is [-0.26 0.28]
Classification of emotion vs. other
ROC_PLOT Output: Single-interval, Optimal balanced error rate
Threshold: 0.04 Sens: 57% CI(47%-67%) Spec: 57% CI(50%-64%) PPV: 40% CI(31%-48%) Nonparametric AUC: 0.57 Parametric d_a: 0.26 Accuracy: 57% +- 3.0% (SE), P = 0.032976
Threshold for cosine_sim = 0.0427, Cohen's d(emo vs no) = 0.26, Bootstrapped CI is [0.01 0.51]
Cos_Sim_Thresh Cohens_d Sensitivity Specificity PPV
______________ _________ ___________ ___________ _______
Pain 0.039225 -0.25929 0.46667 0.46667 0.30435
Cog Control 0.039912 0.0027512 0.48889 0.48889 0.32353
Neg Emotion 0.042662 0.25647 0.56667 0.56667 0.39535
----------------------------------------------
Stats for individual studies
----------------------------------------------
Col 1: ThermalPain1 Col 2: ThermalPain2 Col 3: VisceralPain1 Col 4: VisceralPain2 Col 5: MechanicalPain1 Col 6: MechanicalPain2 Col 7: Cog WM1 Col 8: Cog WM2 Col 9: Cog Inhib1 Col 10: Cog Inhib2 Col 11: Cog RespSel1 Col 12: Cog RespSel2 Col 13: Emotion Aversiveimages1 Col 14: Emotion Aversiveimages2 Col 15: Emotion Rejection1 Col 16: Emotion VicariousPain2 Col 17: Emotion AversiveSound1 Col 18: Emotion AversiveSound2
---------------------------------------------
Tests of column means against zero
---------------------------------------------
Name Mean_Value Std_Error T P Cohens_d
___________________________ __________ _________ _______ __________ ________
{'ThermalPain1' } 0.04189 0.012472 3.3586 0.0046827 0.8672
{'ThermalPain2' } 0.025717 0.014127 1.8204 0.090135 0.47003
{'VisceralPain1' } 0.023952 0.014001 1.7107 0.1092 0.44169
{'VisceralPain2' } 0.015137 0.018477 0.81923 0.42638 0.21152
{'MechanicalPain1' } 0.025621 0.021969 1.1663 0.26299 0.30113
{'MechanicalPain2' } 0.037394 0.013658 2.7379 0.01602 0.70693
{'Cog WM1' } 0.10199 0.016602 6.1434 2.5474e-05 1.5862
{'Cog WM2' } 0.095974 0.01035 9.2731 2.359e-07 2.3943
{'Cog Inhib1' } 0.018087 0.0087099 2.0766 0.056738 0.53616
{'Cog Inhib2' } 0.033083 0.014329 2.3089 0.036728 0.59614
{'Cog RespSel1' } 0.023929 0.019583 1.2219 0.24192 0.31549
{'Cog RespSel2' } -0.033707 0.015111 -2.2306 0.042579 -0.57595
{'Emotion Aversiveimages1'} 0.043905 0.015278 2.8737 0.012263 0.74198
{'Emotion Aversiveimages2'} 0.057978 0.016536 3.5062 0.0034916 0.90531
{'Emotion Rejection1' } 0.058618 0.0089148 6.5753 1.2384e-05 1.6977
{'Emotion VicariousPain2' } 0.0093978 0.017553 0.53541 0.60077 0.13824
{'Emotion AversiveSound1' } 0.052505 0.016395 3.2025 0.0063888 0.82688
{'Emotion AversiveSound2' } 0.084391 0.020375 4.1418 0.00099733 1.0694
Visualization of images
f = 'dACC_pain_pattern_wani_121416.nii';
obj = fmri_data(f);
Using default mask: /Users/torwager/Documents/GitHub/CanlabCore/CanlabCore/canlab_canonical_brains/Canonical_brains_surfaces/brainmask_canlab.nii
Defaults settings have been modified by file(s):
/Users/torwager/Dropbox (Dartmouth College)/Matlab_code_external/spm12/spm_my_defaults.m
Modified fields: stats
sampleto = 'dACC_pain_pattern_wani_121416.nii'
Defaults settings have been modified by file(s):
/Users/torwager/Dropbox (Dartmouth College)/Matlab_code_external/spm12/spm_my_defaults.m
Modified fields: stats
loading mask. mapping volumes.
checking that dimensions and voxel sizes of volumes are the same.
Pre-allocating data array. Needed: 177488 bytes
Loading image number: 1
Elapsed time is 0.036108 seconds.
Image names entered, but fullpath attribute is empty. Getting path info.
Number of unique values in dataset: 247 Bit rate: 7.95 bits
Warning: Number of unique values in dataset is low, indicating possible restriction of bit rate. For comparison, Int16 has 65,536 unique values
create_figure('montage'); axis off
montage(obj);
Setting up fmridisplay objects
sagittal montage: 37 voxels displayed, 209 not displayed on these slices
sagittal montage: 62 voxels displayed, 184 not displayed on these slices
sagittal montage: 51 voxels displayed, 195 not displayed on these slices
axial montage: 79 voxels displayed, 167 not displayed on these slices
axial montage: 65 voxels displayed, 181 not displayed on these slices
create_figure('montage2'); axis off
montage(region(obj), 'regioncenters', 'colormap');
Grouping contiguous voxels: 1 regions
f = 'dACC_task_pattern_wani_121416.nii';
obj = fmri_data(f);
Using default mask: /Users/torwager/Documents/GitHub/CanlabCore/CanlabCore/canlab_canonical_brains/Canonical_brains_surfaces/brainmask_canlab.nii
Defaults settings have been modified by file(s):
/Users/torwager/Dropbox (Dartmouth College)/Matlab_code_external/spm12/spm_my_defaults.m
Modified fields: stats
sampleto = 'dACC_task_pattern_wani_121416.nii'
Defaults settings have been modified by file(s):
/Users/torwager/Dropbox (Dartmouth College)/Matlab_code_external/spm12/spm_my_defaults.m
Modified fields: stats
loading mask. mapping volumes.
checking that dimensions and voxel sizes of volumes are the same.
Pre-allocating data array. Needed: 177488 bytes
Loading image number: 1
Elapsed time is 0.028957 seconds.
Image names entered, but fullpath attribute is empty. Getting path info.
Number of unique values in dataset: 247 Bit rate: 7.95 bits
Warning: Number of unique values in dataset is low, indicating possible restriction of bit rate. For comparison, Int16 has 65,536 unique values
create_figure('montage'); axis off
montage(obj);
Setting up fmridisplay objects
sagittal montage: 37 voxels displayed, 209 not displayed on these slices
sagittal montage: 62 voxels displayed, 184 not displayed on these slices
sagittal montage: 51 voxels displayed, 195 not displayed on these slices
axial montage: 79 voxels displayed, 167 not displayed on these slices
axial montage: 65 voxels displayed, 181 not displayed on these slices
create_figure('montage2'); axis off
montage(region(obj), 'regioncenters', 'colormap');
Grouping contiguous voxels: 1 regions
f = 'stroop_pattern_wani_121416.nii';
obj = fmri_data(f);
Using default mask: /Users/torwager/Documents/GitHub/CanlabCore/CanlabCore/canlab_canonical_brains/Canonical_brains_surfaces/brainmask_canlab.nii
Defaults settings have been modified by file(s):
/Users/torwager/Dropbox (Dartmouth College)/Matlab_code_external/spm12/spm_my_defaults.m
Modified fields: stats
sampleto = 'stroop_pattern_wani_121416.nii'
Defaults settings have been modified by file(s):
/Users/torwager/Dropbox (Dartmouth College)/Matlab_code_external/spm12/spm_my_defaults.m
Modified fields: stats
loading mask. mapping volumes.
checking that dimensions and voxel sizes of volumes are the same.
Pre-allocating data array. Needed: 177488 bytes
Loading image number: 1
Elapsed time is 0.032031 seconds.
Image names entered, but fullpath attribute is empty. Getting path info.
Number of unique values in dataset: 1878 Bit rate: 10.87 bits
create_figure('montage'); axis off
montage(obj);
Setting up fmridisplay objects
sagittal montage: 65 voxels displayed, 1812 not displayed on these slices
sagittal montage: 68 voxels displayed, 1809 not displayed on these slices
sagittal montage: 72 voxels displayed, 1805 not displayed on these slices
axial montage: 500 voxels displayed, 1377 not displayed on these slices
axial montage: 556 voxels displayed, 1321 not displayed on these slices
create_figure('montage2'); axis off
montage(region(obj), 'regioncenters', 'colormap');
Grouping contiguous voxels: 61 regions
